Microsoft Advertising has quietly updated its help documentation to confirm that Predictive Matching is being moved into Search term matching, housed within AI Max. It is a consolidation rather than a removal - but the framing matters. Predictive Matching, which extended reach beyond exact keyword intent using audience signals, is now being folded into a broader AI-driven targeting layer. The implication is that the platform is tidying up its AI features into a single structure, not offering them as parallel choices.
That kind of structural change rarely stays theoretical. When a platform reorganises how its targeting options are labelled and grouped, it usually signals which direction the product is heading. AI Max is where Microsoft wants advertisers. Predictive Matching as a standalone concept is being absorbed into that.
What Predictive Matching Actually Did
Predictive Matching allowed Microsoft Ads to serve ads on queries that went beyond the advertiser's keyword list, using audience and contextual signals to infer relevance. In practice, it functioned similarly to broad match on Google - extending reach in ways that could be productive or wasteful, depending on how well conversion signals were feeding back into the system.
The core tension with any predictive or AI-driven matching has always been visibility. Advertisers could see what queries triggered their ads in search term reports, but the logic behind why a query was selected remained opaque. That opacity is not unique to Microsoft - it is a systemic characteristic of AI-driven query expansion across platforms.
Moving this capability under the Search term matching umbrella within AI Max does not resolve that opacity. What it does is consolidate the controls into a single interface, which could make it easier for advertisers to manage - or easier for the platform to expand reach under a more unified permission.
The Parallel with Google Is Worth Taking Seriously
Google has been running a similar playbook. AI Max for Search on Google consolidates URL expansion, text customisation, and broad match-style query matching under one campaign-level toggle. Broad match has been repositioned from a targeting option into a Smart Bidding companion. Dynamic Search Ads are being retired in favour of AI Max. The direction is unmistakable: traditional keyword-level control is giving way to intent interpretation at the campaign level.
Microsoft is following the same architecture. Predictive Matching sat awkwardly as a separate concept when AI Max was introduced. Folding it in makes the product more coherent from a platform perspective. From an advertiser perspective, the question is whether that coherence comes at the cost of granularity.
On Google, advertisers managing lead generation campaigns have spent the last two years working out which AI Max controls actually limit unwanted expansion - negative keywords, placement exclusions, URL rules - and which ones are largely cosmetic. Microsoft advertisers will face the same process. The settings exist, but understanding which levers have real effect requires testing, not documentation.
What Changes for Campaign Structure
If you are running Microsoft Ads campaigns with Predictive Matching enabled, the transition into AI Max's Search term matching is the immediate practical consideration. The exact migration path was noted in Microsoft's help documentation updates, so checking those pages directly is the right starting point rather than relying on assumptions about how settings will carry over.
The wider structural question is whether to opt into AI Max at the campaign level at all. AI Max on Microsoft gives access to Search term matching, URL expansion, and ad text customisation. These features come as a package. If you need fine-grained control over where your ads appear and what they say - which most lead generation advertisers do - you need to audit the controls available within AI Max before expanding its scope, not after.
Negative keyword lists remain the most reliable constraint available inside these AI-driven frameworks on both platforms. Building and maintaining them is not optional. If Predictive Matching was already pulling in irrelevant queries for your account, those same query patterns will likely surface through Search term matching in AI Max unless negatives are in place to block them.
Conversion Signal Quality Is the Real Variable
Whether AI-driven query expansion produces good results or expensive noise comes down to what signals the platform is optimising against. A campaign optimising toward a low-quality conversion event - a page visit, a form impression - will expand toward traffic that satisfies that signal regardless of commercial intent. A campaign feeding qualified lead data back into the system has a materially better chance of the expansion being directionally right.
This holds on Microsoft just as it does on Google. If Predictive Matching was delivering poor lead quality in your account, simply transitioning to Search term matching within AI Max will not fix that. The fix sits upstream - in what conversion event the campaign is optimising toward, and whether offline conversion data or lead quality signals are being fed back to inform the bidding.
Microsoft does support offline conversion import, and it supports conversion value rules to weight leads differently. Advertisers who have not set these up are leaving the AI to guess what good looks like. That is a structural weakness that a platform migration will not correct.
How to Approach the Transition Practically
Start with your search term reports. Before the transition fully beds in, pull a review of what Predictive Matching was surfacing. Identify any query patterns that were consistently irrelevant or generating low-quality traffic. Those become the foundation of your negative keyword additions going into the AI Max structure.
Then look at your conversion setup. If your Microsoft Ads campaigns are only tracking website conversions and not pulling in any lead quality or offline data, address that before assuming AI Max will improve performance. The matching logic is only as useful as the outcome data it has to work with.
Finally, treat the AI Max transition as a campaign audit opportunity. Settings that made sense when Predictive Matching was a separate toggle may need revisiting in a consolidated structure. Budget allocation, bid strategies, and URL expansion settings are all worth checking explicitly rather than allowing them to carry over by default. Defaults in AI-driven campaigns tend to favour reach. Most lead generation accounts are better served by starting conservative and expanding deliberately.